JP2003348382A - Contour emphasizing circuit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a contour emphasizing circuit in which high quality edge emphasis is realized. <P>SOLUTION: Based upon a weighting coefficient (k) imparted from a coefficient calculation circuit 36, a weighting adder circuit 32 applies weighting addition to raw image data G and data Gm after the relevant raw image data G are filtered by a median filter 30. Based upon the weighting added result Ga, an aperture signal generating circuit 38 generates an aperture sinal A. The aperture signal A is superimposed on Y data outputted from a signal processing circuit 22 in an adder circuit 26 such that an edge component of the relevant Y data is emphasized. <P>COPYRIGHT: (C)2004,JPO

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a contour emphasizing circuit, and more particularly to a contour enhancing circuit, for example, applied to a digital camera, and a first pixel signal output from a photographing means and a first pixel signal obtained by performing a filtering process on the first pixel signal. The present invention relates to an outline emphasis circuit that generates an outline emphasis signal by weighted addition with a two-pixel signal.

[0002]

2. Description of the Related Art One of image processing techniques in a digital camera is edge emphasis that sharpens an outline of an image. This is a technique of extracting an edge (high-frequency) component of an image signal output from an image sensor, adding an aperture gain to the extracted edge component to generate an aperture signal, and superimposing the aperture signal on the image signal. . In order to extract an edge component of an image signal in such edge enhancement, conventionally, an edge extraction process has been directly performed on the image signal.

[0003]

However, according to the above-mentioned prior art, when a noise component having a relatively high frequency is included in the image signal, the noise component is also extracted together with the edge component. As a result, an aperture gain is given to the extracted noise component, so that not only the edge component but also the noise component is emphasized, resulting in a problem that the image quality is deteriorated.

Therefore, a main object of the present invention is to provide a contour emphasizing circuit capable of realizing high-quality edge emphasis.

[0005]

According to the present invention, an outline emphasis signal is obtained by weighted addition of a first pixel signal output from a photographing means and a second pixel signal obtained by performing a filtering process on the first pixel signal. Detecting means for detecting a level change amount of the first pixel signal in the contour emphasizing circuit for generating
And a weight control means for increasing the weight of the first pixel signal as the level change increases.

[0006]

According to the present invention, an outline emphasis signal is generated by weighted addition of the first pixel signal output from the photographing means and the second pixel signal obtained by performing a filtering process on the first pixel signal. Here, for example, when a contour portion of a subject exists at a certain pixel position, the first pixel signal corresponding to the pixel shows a relatively large level change. on the other hand,
The first pixel signal of a pixel located in a substantially flat part of the object shows only a relatively small level change. This first
The level change amount of the pixel signal is detected by the detecting means. The weighting control circuit increases the weighting amount for the first pixel signal as the detected level change amount increases. That is, the contour emphasis signal of the contour portion of the subject is generated by weighted addition in which the so-called original first pixel signal output from the photographing means is weighted heavily. Then, the outline emphasis signal of the substantially flat portion of the subject is generated by weighted addition in which the second image signal in which the noise component has been reduced by the filter processing is heavily weighted.

[0007] Median filter processing can be used as the filter processing.

Further, it is desirable that the first pixel signal is a signal having green color information.

[0009] Further, there may be further provided generating means for generating a luminance signal based on the first pixel signal, and adding means for adding a contour emphasis signal to the luminance signal.

[0010]

According to the present invention, the contour emphasizing signal of the contour portion of the subject is generated by weighted addition in which the original first pixel signal output from the photographing means is weighted greatly. Can further enhance the sharpness. On the other hand, since the contour emphasis signal of the substantially flat portion of the subject is generated by weighted addition in which the second pixel signal in which the noise component is reduced from the first pixel signal is greatly weighted, the influence of the noise component in the flat portion is obtained. Is reduced. Therefore, it is possible to realize higher-quality edge enhancement than before.

The above objects, other objects, features and advantages of the present invention will become more apparent from the following detailed description of embodiments with reference to the drawings.

[0012]

Referring to FIG. 1, a digital camera 10 of this embodiment has a CCD (Charge Coupled Device) type image sensor 14 into which an optical image of a subject is incident through an optical lens 12. The image sensor 14
According to a timing signal supplied from a TG (Timing Generator) 16, an optical image of a subject is exposed,
An image signal corresponding to the charge accumulated by the exposure is output.

An image signal output from the image sensor 14 is a CDS (Correlated Double Sampling) / AGC signal.
(Automatic Gain Control) circuit 18, where the signal is subjected to correlated double sampling processing and amplification processing, and then input to A / D conversion circuit 20. The A / D conversion circuit 20 converts the input image signal into image data which is a digital signal, and inputs the converted image data to the signal processing circuit 22. The signal processing circuit 22 performs color separation, white balance adjustment, gamma correction, YUV
After performing predetermined processing such as conversion, Y (luminance) data, U (blue difference) data, and V (red difference) data generated by the YUV conversion are input to a video encoder and a compression circuit (not shown). As a result, a through image of the subject is finally displayed on a screen of a liquid crystal monitor (not shown), and a compressed image file is recorded on a memory card (not shown).

Further, the digital camera 1 of this embodiment
0 has an edge emphasis circuit 24. The image data converted by the A / D conversion circuit 20 is input to the edge enhancement circuit 24, and the edge enhancement circuit 24 generates an aperture signal A from the input image data. The generated aperture signal A is input to the addition circuit 26, where it is added (superimposed) to the Y data output from the signal processing circuit 22. Thus, the edge component of the Y data is enhanced, that is, edge enhancement is realized.

More specifically, as shown in FIG. 2, the edge emphasizing circuit 24 includes a median (median) filter 30, a weighting and adding circuit 32, an edge level detecting circuit 34, a coefficient calculating circuit 36, and an aperture signal generating circuit 38. including. And, among these, the median filter 30,
Weight addition circuit 32 and edge level detection circuit 34
, The image data after the conversion by the A / D conversion circuit 20 is input.

The image data converted by the A / D conversion circuit 20 is, as shown in FIG.
d), a plurality of pixel data having one of G (Green) and B (Blue) color components. The color components of these pixel data correspond to the color elements of a primary color filter (not shown) provided on the light receiving surface of the image sensor 14.

That is, FIG. 3 conceptually shows the color components of each pixel data in correspondence with the arrangement of each pixel (light receiving element) of the image sensor 14. When attention is paid to a certain vertical column, In the columns, pixel data of the R component and pixel data of the G component are alternately arranged. Then, in another vertical column adjacent to the vertical column, pixel data of G component and B
The pixel data of the component are alternately arranged. Further, FIG.
As shown by the thick line frame 50, 4 of arbitrary 2 rows × 2 columns
Focusing on one pixel, pixel data of the G component is arranged on one diagonal of the frame 50, and pixel data of the R component and pixel data of the B component are arranged on the other diagonal.

Returning to FIG. 2, the median filter 30
Performs a 3 × 3 filter process on pixel data of the G component among a plurality of pixel data included in the image data. That is, as shown in FIG. 4, the median filter 30 sets an arbitrary G pixel (a pixel to which a G color element is assigned) as a target pixel G0, and sets the target pixel G0 and four G pixels surrounding the target pixel G0. The median value of each of the pixel data G1 to G4, that is, the third largest pixel data from the top (or the bottom) is specified. Then, the specified pixel data is set as the pixel data of the target pixel G0. This filtering process is performed on the pixel data of all the G pixels other than the G pixels existing at the edge of the image sensor 14. The pixel data of the G pixel existing at the edge of the image sensor 14 is maintained as it is.

The filtering process by the median filter 30 reduces noise components included in the image data (G component pixel data). The image data Gm after the filter processing is input to the weighting addition circuit 32.

The weighting and adding circuit 32 outputs the filtered image data G input from the median filter 30.
m, image data input from the A / D conversion circuit 20,
A weighted addition represented by the following equation 1 is performed on so-called raw image data.

[0021]

Ga = Gm · k + G · (1−k) Here, Gm is the pixel data of the G component after the filtering process, and G is the pixel data of the G component included in the raw image data. K is a weighting coefficient. This weighting coefficient k is obtained by the following procedure.

First, the edge level detection circuit 34 detects A /
The calculation represented by Expression 2 is performed on the G component pixel data in the raw image data input from the D conversion circuit 20.

[0023]

E = | {(G0-G1) / 2} + {(G0-G
2) / 2｝ + ｛(G0-G3) / 2｝ + ｛(G0-G
4) / 2｝ | = | 2 · G0− (G1 + G2 + G3 + G4) / 2 | Here, G0 is the pixel data of the target pixel G0 in FIG. 4, and G1 to G4 are the four peripherals surrounding the target pixel G0, respectively. This is pixel data of the pixels G1 to G4.

As can be seen from Equation 2, the edge level detection circuit 34 calculates the absolute value of the sum of the values obtained by dividing the difference between the pixel data of the target pixel G0 and the pixel data of each of the peripheral pixels G1 to G4 by two. . Here, for example, if there is a contour portion of the subject at the position of the target pixel G0, the calculation result E by the edge level detection circuit 34 becomes a relatively large value. On the other hand, the target pixel G0 and the peripheral pixels G1 to G
When 4 is located in a substantially flat part of the subject, the calculation result E is a relatively small value, and extremely E = 0. That is, the edge level detection circuit 34 detects the degree of the edge of the subject, in other words, the flatness of the subject.

The result of detection by the edge level detection circuit 34, that is, data E representing the degree of an edge, is input to a coefficient calculation circuit 36. The coefficient calculation circuit 36 calculates the above-mentioned weight coefficient k based on the input data E. The formula for calculating the weighting coefficient k is expressed by the following equation (3).

[0026]

K = α; E ≦ Th k = α− (E−Th) · β; E> Th Here, α, β and Th are constant values, respectively.

FIG. 5 is a graph showing the relationship between the data E expressed by the equation (3) and the weighting coefficient k. That is, when the value of the data E is equal to or less than a certain threshold Th,
The weighting coefficient k has a constant value α. On the other hand, in the range where the value of the data E exceeds the threshold value Th, the weighting coefficient k is smaller than α and becomes smaller as the value of the data E becomes larger, that is, as the degree of the edge of the subject becomes larger.

The weighting coefficient k thus calculated
Is given to the weighting addition circuit 32. The weighting and adding circuit 32 performs the above-described weighting and adding of Equation 1 using the given weighting coefficient k. According to Equation 1, the larger the weighting coefficient k, that is, the smaller the degree of the edge, the larger the ratio of the weighting to the image data (Gm) after the filtering process by the median filter 30. On the other hand, the smaller the weighting coefficient k, that is, the greater the degree of the edge, the greater the ratio of weighting to the raw image data (G). The data Ga after the weighting and addition by the weighting and adding circuit 32 is input to the aperture signal generating circuit 38.

The aperture signal generation circuit 38 generates an aperture signal A by extracting an edge component of the input data Ga and applying an aperture gain to the extracted edge component. A series of processes by the aperture signal generation circuit 38 is realized by a known technique. The generated aperture signal A is added to the addition circuit 26 as described above.
Is superimposed on the Y data output from the signal processing circuit 22, thereby enhancing the edge component of the Y data.

As can be seen from the above description, in this embodiment, the aperture signal A is generated based on the weighted addition result Ga of the raw image data (G) and the image data (Gm) subjected to the filtering process by the median filter 30. Is done.
The aperture signal A to be superimposed on the contour of the subject is generated based on the weighted addition result Ga in which the weighting ratio for the raw image data (G) is increased, so that the contour is further sharpened. Can be. On the other hand, the aperture signal A superimposed on the substantially flat portion of the subject is generated based on the weighted addition result Ga obtained by increasing the weighting ratio with respect to the filtered image data (Gm). The effect of noise components can be reduced. That is, high-quality edge enhancement can be realized.

In this embodiment, the case where the present invention is applied to the digital camera 10 has been described. However, the present invention is not limited to this. That is, the present invention may be applied to various devices that handle image signals, such as video cameras and video switchers.

Although the aperture signal A is generated based on the G component image data, the aperture signal A may be generated based on the R component or B component image data. However, since the luminance component that has the greatest effect on the edge enhancement effect is included most in the G component image data, generating the aperture signal A based on the G component image data provides the largest edge enhancement effect. Can be obtained.

Further, in place of the median filter 30, another non-linear filter such as a maximum value filter, a minimum value filter, an average value filter, and a steep filter may be used. In addition, a linear filter may be used instead of the nonlinear filter.

Although the edge level detection circuit 34 calculates the degree E of the edge based on the above-described equation 2, the degree E of the edge may be calculated by another arithmetic expression. The relationship between the degree E of the edge and the weighting coefficient E is not limited to the relationship shown in FIG.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of the present invention.

FIG. 2 is a block diagram illustrating a detailed configuration of an edge enhancement circuit in FIG. 1;

FIG. 3 is an illustrative view conceptually showing a color component of pixel data included in image data in the embodiment in FIG. 1;

FIG. 4 is an illustrative view for explaining processing contents of an image signal in the edge enhancement circuit of FIG. 2;

FIG. 5 is a graph showing a relationship between the degree of an edge and a weighting coefficient in the embodiment of FIG. 1;

[Explanation of symbols]

10 Digital camera 22 ... Signal processing circuit 24 ... Edge emphasis circuit 26 ... Addition circuit 30 ... Median filter 32 ... weighted addition circuit 34 ... Edge level detection circuit 36 ... Coefficient calculation circuit 38 ... Aperture signal generation circuit

Claims (4)

[Claims] An edge enhancement circuit for generating an edge enhancement signal by weighted addition of a first pixel signal output from a photographing means and a second pixel signal obtained by performing a filtering process on the first pixel signal, Detecting means for detecting a level change amount of the first pixel signal;
And a weighting control means for increasing a weighting amount for the first pixel signal as the level change amount increases. 2. The contour emphasizing circuit according to claim 1, wherein said filtering is median filtering. 3. The contour emphasizing circuit according to claim 1, wherein said first pixel signal is a signal having green color information. 4. The image processing apparatus according to claim 1, further comprising: generating means for generating a luminance signal based on said first pixel signal; and adding means for adding said contour enhancement signal to said luminance signal. Edge enhancement circuit.